Machine Learning Basics Quiz — Quiz

Question 1 of 10

Predicting house prices is a…

aClassification taskbClustering taskcRegression taskdReinforcement task

Question 2 of 10

Why split data into train and test sets?

aTo evaluate on unseen databTo train fastercTo reduce featuresdTo balance classes

Question 3 of 10

Model performs great on train, poorly on test. This is…

aUnderfittingbRegularisationcNormalisationdOverfitting

Question 4 of 10

Best metric for rare fraud detection?

aAccuracybRecall / PR-AUCcR²dMAE

Question 5 of 10

K-Means is an example of…

aUnsupervised learningbSupervised learningcReinforcement learningdDeep learning

Question 6 of 10

What does `random_state=42` do here?

X_tr, X_te, y_tr, y_te = train_test_split(
    X, y, test_size=0.2, random_state=42)

aIt uses 42% of the rows for testingbIt trains 42 modelscIt makes the model more accuratedIt fixes the shuffling seed so the same split is produced on every run

Question 7 of 10

Precision answers which question?

aHow many of the actual positives did the model find?bOf the cases predicted positive, how many really are positive?cHow close are predictions to the average?dHow quickly does the model train?

Question 8 of 10

A `city` column holds Pune, Delhi and Mumbai with no natural order. Which encoding fits?

aLabel encoding as 0, 1, 2bDropping the columncOne-hot encodingdStandard scaling

Question 9 of 10

Why put the scaler inside the Pipeline instead of scaling X once beforehand?

model = Pipeline([('scale', StandardScaler()), ('clf', LogisticRegression())])
scores = cross_val_score(model, X, y, cv=5)

aEach fold fits the scaler on its training part only, so validation statistics do not leak into trainingbStandardScaler cannot be called on its owncIt makes cross_val_score run fasterdPipelines are required for logistic regression

Question 10 of 10

As a model is made more flexible (deeper trees, more features), what usually happens?

aBoth bias and variance fallbBias falls while variance risescBias rises while variance fallsdNeither changes